planet: 5 sources fail scraping (dead/moved feeds, not timeouts)

[cuihtlauac]

Summary

Planet scraper (while validating the timeout/retry work in
#3752) surfaced 5 sources in data/planet-sources.yml that fail for reasons
unrelated to timeouts — they were already failing before that change and
still fail after. They clutter every scrape report and should be fixed or
retired.

Failing sources

id	name	url	error	user
mirage	MirageOS	https://mirage.io/feed.xml	Failure("Neither Atom nor RSS2 feed")	@hannesm
ocaml-book	OCaml Book	http://ocaml-book.com/blog?format=rss	Failure("Neither Atom nor RSS2 feed")	@johnwhitington
patricoferris	Patrick Ferris' OCaml Blog	https://patrick.sirref.org/ocaml-blog/atom.xml	Failure("Neither Atom nor RSS2 feed")	@patricoferris
dinosaure	Romain Calascibetta	https://blog.osau.re/feed.xml	DNS/name resolution failed	@dinosaure
emilpriver	Emil Privér	https://priver.dev/tags/ocaml/index.xml	Http.Status_unhandled("404 Not Found")	@emilpriver

Notes

• The three "Neither Atom nor RSS2 feed" cases mean the URL returns
  something that isn't a parseable Atom/RSS feed (moved feed, HTML error page,
  or a format river/syndic doesn't accept). Each needs the feed URL rechecked.
• dinosaure — blog.osau.re does not resolve; the blog appears to have moved
  or gone offline. Needs a new URL or removal.
• emilpriver — the tag feed 404s; likely a changed path.

Suggested action

For each: find the current working feed URL and update it, or mark the entry
disabled: true (or remove it) if the blog is gone. Independent of #3752, which
only addresses the timeout class of failures.

[cuihtlauac]

@dinosaure, @patricoferris, @emilpriver, @johnwhitington, @hannesm : Any views on this?

[johnwhitington]

Mine is gone. The link to it can be deleted. Thanks!

[emilpriver]

Mine is gone. The link to it can be deleted. Thanks!

same!

[hannesm]

I don't know how I end up in your list here. FWIW, there is an atom XML returned from "mirage.io/feed.xml". According to the w3 feed validator, it has a validation failure since <p><img src="/img/blog-VPNKit.png" alt="Image showing the modular structure of vpnkit" ></p> is not closing the img properly. I opened dbuenzli/cmarkit#32 with the hope that will fix the issue.

Now, the mirage core team should care about that, rather than me individually //cc @samoht @avsm @talex5 @dinosaure @palainp @MagnusS @TheLortex @djs55

[hannesm]

So, the MirageOS feed was fixed by @reynir in mirage/mirage-www#868). Now the validator.w3.org is fine (it has some recommendations). Not sure what your feed import utility does - but it should behave better now. (You can back out your^Wclaude's 8a3cab0).